RUS  ENG
Full version
JOURNALS // Sistemy i Sredstva Informatiki [Systems and Means of Informatics] // Archive

Sistemy i Sredstva Inform., 2022 Volume 32, Issue 1, Pages 114–125 (Mi ssi817)

This article is cited in 3 papers

Database of hierarchies of logical-semantic relations: Architecture

A. A. Durnovoa, O. Yu. Inkovaab, N. A. Popkovaa

a Federal Research Center "Computer Science and Control" of the Russian Academy of Sciences, 44-2 Vavilov Str., Moscow 119333, Russian Federation
b University of Geneva, 22 Bd des Philosophes, CH-1205 Geneva 4, Switzerland

Abstract: The paper presents the architecture of a new linguistic resource: the Supracorpora database reflecting hierarchies of logical-semantic relations that ensure the text coherence. Annotations in the database are in the form of trees or, rather, arborescences, where vertices contain data while edges represent the subordination between vertices. Each tree vertex corresponds to either a context or a connector; the connectors in the text are marked. The authors describe the relationships between the database tables and trees as well as their properties. The paper also demonstrates the differences between this new resource and the existing ones, in particular, the graphs of rhetorical relations created within the framework of the rhetorical structure theory: the ability to store data, to modify the annotated contexts, to work with empty contexts, and to reflect previous states of all the trees.

Keywords: supracorpora database, corpus of texts' annotation, graph, discourse relations, connector.

Received: 14.02.2022

DOI: 10.14357/08696527220111



© Steklov Math. Inst. of RAS, 2024